The hippocampus is a brain structure that is critical for normal learning and memory functions. For example, one of the first brain regions to begin deterioration in Alzheimer's Disease is the entorhinal cortex, the key processing stage between the neocortex and the hippocampus proper. This degeneration correlates with the memory deficits that are the first cognitive symptoms of the disease. To understand why hippocampal damage causes such severe memory deficits, it is necessary to understand the basic computational functions of this brain region. The first stage of hippocampal processing is the dentate gyrus (DG), which receives major inputs from the entorhinal cortex and projects to the CA3 region, where hippocampus-dependent associative memories are thought to be stored. It has long been hypothesized that the DG preprocesses the data from the entorhinal cortex by performing a pattern separation function-creating output patterns of neural activity that are less similar to each other than the entorhinal input patterns. This operation reduces interference when these patterns are stored as memories in CA3. Tests of this theory have been equivocal, however, in part because the DG contains a number of excitatory cell types that makes it difficult to determine what part of the circuitry is involved in the pattern separation function. These cell types are well-characterized in terms of anatomy and cellular physiology, but their firing correlates in freely behaving animals are not understood. The specific aims of this project are to identify the different cell types in the DG (mature granule cells, new adult-born granule cells, and mossy cells) in freely moving animals and to test the role of each type in pattern separation using classic tests of the place-cell remapping phenomenon. A novel combination of high-density neural recording probes, optogenetics, and juxtacellular labeling will be used to (a) record the spatial firing properties of DG neurons as rats explore different environments and (b) assign these properties to the specific cell types. The mature granule cells are hypothesized to perform the pattern separation function ascribed to them by the classic computational theories, reflected in their ultra-sparse firing and their extreme activity-based discrimination of different environments. In contrast, the newborn granule cells and mossy cells are hypothesized also to take part in the pattern separation computation, but with the use of a different coding strategy than the mature granule cells, firing promiscuously in different environments but with different rates and patterns of spatial activity in each environment. The results of these experiments will provide crucial knowledge of the physiological roles of each of these cell types, and will greatly enhance the ability of future work to test the many questions about the roles of the DG and adult neurogenesis on learning and memory function.